ChatGPT’s viral caricature trend proves how well it knows you — and it’s freaking people out

There’s a new AI trend that’s making us wonder if it knows us a little too well.

Worldwide, people have taken to asking the AI platform, ChatGPT, to create a caricature of them based on what it already knows about them.

This information will usually come from their chat history, but the accuracy can be pretty astounding.

Users on X have been sharing their caricatures, pointing to how well they’ve been captured by the AI tool.

Many will have things in the background that round out the picture with extra detail: books with uncanny titles, screens with graphs or editing software depicted on them, flags, and even fantasy football stats can be seen in the background of the caricature images.

While similar images can be made in apps like Cartoonify, ChatGPT offers the twist of proving how well it “knows” you based on what you’ve previously asked or searched on the platform.

However, the excitement around the trend feels a little paradoxical, given the concern many hold that AI “knows too much.”

If it knows us better than we know ourselves, what does that say for our privacy, and what AI platforms choose to do with the data we provide to them?

While it seems like a harmless trend, AI platforms aren’t bound by confidentiality agreements the way doctors, lawyers, and psychologists are.

David Grover, Senior Director of Cyber Initiatives at Baylor University, told KWTX that once you upload information, you lose control of it.

“In general, all of them are going to save your information, and that information is going to go into some kind of storage, and at that point, you don’t really know what that company is doing with your information,” he explained.

“You need to be careful about any image that you put in and anything that you put online because that becomes who you are, and the more we get into the digital world, the more challenging it’s going to become to protect.”